Download Machine Learning Assignment: Introduction to Machine Learning Concepts and Techniques and more Assignments Machine Learning in PDF only on Docsity!
MACHINE LEARNING
ASSIGNMENT 1 Nae: NISHA KAUSH1K Rlno: 19PG CSo
Q hat^ du^ you^ mean^ by^ Machine^ Leasunin^? Dleunuate b^ atunun^ Supauuise^ d^ and
unsu peLLiSe^ d^ aunin9.
ms Machine^ leauninq^ enablu^8 a^ mochi^ ne^ toto
auutomaically lean^ Mom^ d^ at^ a,^ impAou
p amance Hom exp euence^ 3,^ and^ pHod^
ct thinqs wwthout^ being^ explicitl^ PuDammed Machine Lewnninq^ a^ 8ub&tt^ a^ AI with the^ hulþ o sampla^ histori^ cal^ d (^) ata, which
13 nouin^ as^ Traininq^ Data.^ machine^ uani^ ng
alguthms b^ ulld^ a^ " Mathem^ ati^ cal^ Modil"^ at heus un^ makinq^ pas^ di^ du^ on^ S (^) a dsision3 (^) ui th dut
ben expuatly^ pDammed.^
Th mau
uu (^) p0w di^ thu^ in^ @umati^ oD,^
Hhu higher wiu
be ho^ pe m an^ ce. Machine leauninq^ Can^ be^ classiie^ d^ as (^) b elow :
CA Sup used lsaHninq
B) UnS^ upeuis ed Leainin
LA) Sup^ uised^ Leastninq^ a^ methud^ in^ uhich^ ul pHOids Sample lablo d dat a to Hhu 3ygkm in Odtxi wain it, and an th basie,ib psudicts h 0up ut (^8) unsupesuuised Leauninqamthod n uuhich a mochine eans uitmd'ur any 8up eui si on. The lainin 9 pOuids uüth thi tt a d ala itha has (^) not bean Jab tud,cla ssitd ocateqavized anà tha alqorithm needs o akt an that dalta wi hout any8upeuuision
Supouised Le.auninq Dtained usinq Labed data Lb aka direet jeedback h cherk iit psudicing C0uttolp nob. pudicts He ourput
Un Supeuui sed laaxninq usinq unlabelad data does (^) not (^) takA anu
La)
dbackk
Jinds thu^ hidden^ patern in data ony in put data io p.suoui dsd o tho model
c
Lo)inpur dato o proLidod tD th mddsl alonq uith h uBbut ce) need8^ &upauiuon^ D ain h modal. CF Cateqaized in Classii- cation and^ Requ 8ion
pHoblems.
does not nard oany
Supauusiom to kain
cLassiied an clusierinq and (^) A'sS o ciati (^) on pJHOblo ms.
B) MetsucR Some comman mettuds used to
Oualllat modsl
ClaSSicatiao melucs =Uhen p.aasAM 1N
classioatian pudiction2 ,th is our type
0utcomes hat couwd occu
alse pD Sin uw talse neq ati^ ws
he3e uaut come1 ast^ aten plottd of a
ConuSidn maUx.
he tH maun^ metucs^ USed^ to^ eUauate^ a
la 83hi cation moda 0U
ACCusiacy io dein ed as tha 7 a C8tt
pudi ction8^ Jor^ "^ hu^ Test^ data
O.cc.uHdcy C0Ocal pwdi^ 9Hud dian^ cuons s
Pu cisiCn io dained as the Haction a ulwant examples (u posius) amo'ng au
th eX ampla3uhich ultu pu dicted" 4+
au
belonq un^ a^ Atan^ cd
w posiiw t fa) se positiuu Mean Ab saluute^ Euok^ tha^ auwaqe^ o^ he
dilunce bw^ 0uqina^ ualu&^ and^ h^
Psudicted
Uau S^ t^ quus us^ H^ measu^ a)^ how^ Jax^ th pwdictions w Juom ha acbualautbut, 2
- what o Stasti cal Leasaning Thegy 2 ex plain
dms Statisucal^ leauning^ heosy^ Jo^ a^ am^ eu^ ok^10
o
Machine Leauning dt^ aung^ om^ ths^ Jiuds^ a StaistcS and^ Juncbonal^ Analy^ 3is.^ Jt^ deal3^ with he (^) oblam a Jinding^ a^ þw^ di^ ctiuu^ un^ ction^ based on data.^ Stattstucal^ Leauninq^ Thooy^ has^ Jud^ to Successul appi^ cati^ on^9 in^ eld^9 8uch^ ag Computer^ u^ S1om,^ 8peech^ ucoqnib^
on an d
bio-in g*m atie9. he (^) qo al^3 aunng^ a^
Uunde&tan ding asd^ psud-
i c^ an.^ keasuninq^ Jals^
unt (^) many at e9^ oeS Un (^) clu dn9 3up wused^ Laasunt^ ng^
unS up euuused
auning onu^ no^ oaLni^ nq^ and (^) oi (^) n| o Cment Janina
- Mtan8 Clusteringdan un supenui sed lear- ningalg.otu hm.Hax daines Hha no.a psus doird cluster 8 that n eed to be suated. hs -meang clus terinq alasuth m maint e 0ms tuuo tasks Dut etmines the best valule 0 K Cente% Doints 0CenDids by an ltest cutiuo pDces HsSiqns each ddta points to its DSest k-Conter ThD& daàa bainb uhich aHo near tD ha pasticu LO Cnte CD.as a clulte he wor King o - moang alq 09uithm u - S entt thi no^ o\ K tD docido thn no, a duste Gi)Seutt random x- paints 0e cenHtbids iit)Assiqn eoth datapoint to their dpSe gt CentDid,
which uüu orm Hhe pudalined Kclustts
Liv)Calcuuate thu vaiance CLnð place a new cenoid
eac clusteK
Lv Reea th iii) 3p (vid (^) any 0n381qnnm ent OceMS (^) g 0D 810 li Finis Th (^) modal ib (^) ad
olse Lvii)
- (^) explain the^ model^ supUS^ EN^ tabon a^ Un^ ea
L ASLon^ ui^ Hh^ on^ Vauable
ine ak^ 9uQHU8^ uon^ do^ ong^ o^ the^ mo8tamOua^ wa^ to dus cube yowe data an d maka bsu dfcu ons en E.
Oveuew o Linea Reqyuion Aqosithm
Ttainin Set
Leuning Alqorithm
Hypathesis Juncion
esbm att d
house (^) Pice The (^) uain in q su^ a Housin^ q lsuce^ in^ td nto^ h Leauning Ho thm. s^ mäin^ job to^ haoduce a untti o0, which^ by^ onuntien^ w^ callad^ h^ for^ hypat he3is). Yau thun use tmat hypathesis uncti on to
Qutpul th estimate bice y bu ghuinqtt the di
a hoU3e^ un^ nput^ c. hg (oc)^ +0,^ (x)^ Th (^) eta's (0 n^ qenetal)^ as^ h Pauametes^ o^ the wun cu on. n abouu^ hyp heSis un^ cion,^ hau^ s^ any^ ans ULatiabla ,i.e.", dat his sUa8an,it io callod Linea RegsusSaon uith ona ya ablu.
qtt batk^ to^ Ahs^ Ofigin (^) a) matu%.^ Mabux (^) Jactosuzation Can bt Used to ds coUU Latent je at Usg. examplu n a CD m mend ation system ik Psume Ne| Ux, ha a ADUP o Usg and a seb a t em s.Giwn hat ta Ch use nau 9Oted 3 om item (^) in Hhu (^) SyStem and uld^ uud^ iko to (^) prediut
how thu users wauU d fate +ha items Hhat th u
not uet ated gu ch a9 UU can mal CO Mm d ati on s D he use n thi g ca3e^ al^ ho^ inamati^ on^ u^ hau^ abauh^ ho e1shn q^ at^ ngs^ Can^ be^ Hou0S^ entéd^
unamatuix.
sSunme w^ hau^5 us3^ and^ 1o^ item3^ and 90tin^9 3 C^ V^ aluus^ ntea^ anqn^ om^ I (^) to (^) 5, 8D Hhu^ matux^ m^ JDok^ So^ mehinq^ Uke^ Hhus
D D3 Dy
U 1 (^03 )
U5 5
